In the Drosophila eye, inner photoreceptors (PRs) R7 and R8 mediate color vision and belong to two subtypes, expressing UV-sensitive Rh3 in R7 and Blue-Rh5 in R8 (p subtype), or UV-Rh4 in R7 and Green-Rh6 in R8 (y subtype). Reminiscent of the stochastic distribution of L and M cones in the human retina, the two subtypes are distributed stochastically, with a p:y ratio of 35:65. We have deciphered the gene regulatory network that controls PR subtype specification. The random choice between p and y fate is controlled by stochastic expression of the transcription factor Spineless in 65% of R7s where it induces yR7/Rh4 fate. We have shown that each spineless allele makes an intrinsic stochastic choice to be expressed; however, inter-chromosomal communication synchronizes expression of the two alleles. We will address whether the mechanisms of stochastic choices can be generalized, and why Drosophila uses a stochastic distribution while other species exhibit highly deterministic PR patterning. We will use transcriptomics and cis- regulatory dissection to continue our investigations of PR subtype specification. Finally, we will investigate how PRs that are stochastically determined find their corresponding targets in the optic lobes to convey color information to the brain. Aim 1. Stochastic vs. deterministic choices in photoreceptor determination. Using the MS2 system, manipulation of the spineless proximal promoter and its epigenetic landscape, we will live image spineless transcription in PRs to understand whether spineless is initially stochastically repressed, or if noisy fluctuations are locked-in at a given time. We will analyze the role of spineless in other species, e.g. in butterflies that have two R7s that make independent stochastic choices, and use CRISPR to knock out spineless in these species. Another fly family, Dolichopodidae (Doli), has eyes with alternate stripes of Green and Blue R8s. We will address the regulation and function of spineless in this deterministic process. We will swap the Doli spineless gene into D. melanogaster to make it deterministic. Finally, we will analyze how the neural network is reconfigured in the visual system of males of many fly species to build an improved motion detector. Aim 2. Cis- and trans-regulatory logic of photoreceptor subtype specification. We have obtained a comprehensive knowledge of the gene regulatory network controlling PR determination and the cis-regulation of Rh promoters. However, PR specification precedes Rh expression by two days. We will therefore analyze how early genes (dpr11, dve, warts and melted) are expressed in subtypes of PRs and compare the logic of regulation at several stages for these different genes with functions in PR specification or axon pathfinding. We will compare the early and late transcriptomes of the four types of color as well as polarization PRs and identify potential regulators predicted by our cis-analysis of Rh promoters. Aim 3. How do stochastically determined PRs find their target in the optic lobes? We will investigate whether and how p and y R7 and R8 connect to dedicated subsets of neurons in the medulla and address how these neurons are specified to transmit color information. We will study the role of Ig proteins Dprs and their DIP receptors that allow the matching of PRs and their target in the medulla. Finally, we will study how stochastically determined PRs specify their target neurons or allow survival of pre-existing medulla neurons.